All categories/Translation and localisation
Localisation and release · Level Advanced · Localisation Review Process
Review: who should read the localised text
The translation passed QA and local colleagues say it reads wrong.
Prompt
Localisation is more than words: date formats, currency, faces in images, legal requirements. Without them the text reads foreign.
Task: The translation passed QA and local colleagues say it reads wrong.
[Before you start, do three things]
1. Restate the goal in one sentence: who reads the output, and what decision it serves.
2. If I have not given you the following, ask me. Do not assume:
- language and market
- current review flow
- reviewer background
- time
3. Say how you plan to do it (at most 5 steps) and wait for my go-ahead.
[Go through each of these]
- three layers: accuracy, naturalness, market fit
- native speakers living locally for the third
- a checklist, not 'take a look'
- log disputes with one arbiter
[Process]
1. List every localisable element: text, formats, images, law
2. One rule set per market
3. Check for hard-coded strings
4. Design a verification
5. Say what should stay untranslated
[Output]
Element list, market rules, gap check, verification.
[Do not]
- Never translate only the UI strings
- Do not ignore plurals and gender
- One format never fits all markets
- Never force-translate a brand nameWhat to state before it starts
Leave these unsaid and the model fills them with an assumption. Most 'wrong answers' start here.
- language and market
- current review flow
- reviewer background
- time
Go through each of these
Where this kind of work most often goes wrong. Already in the prompt, so you do not have to remember it.
- three layers: accuracy, naturalness, market fit
- native speakers living locally for the third
- a checklist, not 'take a look'
- log disputes with one arbiter